DESCRIPTION: (Applicant's Abstract) The long-term objective of this research program is to improve radioimmunotherapy (RIT) for human cancers through the use of antibody-pretargeting strategies. The fundamental hypothesis underlying this work is that optimized single-chain Fv molecule (scFv) based proteins will be ideal vehicles for pre-targeted radioimmunotherapy, since the selective targeting advantages of scFv-based molecules are best seen during the terminal phases of in vivo biodistribution. In the current period of support, the applicant has made substantial progress in identifying the relative contributions of antibody size, valence and binding affinity to quantitative and selective tumor retention in murine models. The C6.5 scFv was isolated from a human phage display library panned against the extracellular domain of HER2/neu, which is an important target in breast cancer and other neoplasms. C6.5 affinity variants were generated using chain-shuffling and site-directed mutagenesis to yield a series of scFv with affinities for HER2/neu ranging from 10(-6) - 10(-11) M. Using these molecules, the applicant has determined that the threshold affinity for detectable in vivo tumor targeting in a murine model is 10(-8) M. Increasing the affinity of scFv to more than 10(-9) M does not further improve quantitative, selective tumor retention. Cumulative targeting selectivity, determined by measuring tumor to normal organ area-under-the-curve ratios, does not exceed that seen with larger IgG molecules, although selective tumor targeting by scFv molecules is substantially better during the terminal phases of biodistribution. Increasing valence has more effect than does affinity on quantitative tumor targeting, even when the results are corrected for antibody size. Divalent scFv exhibit profoundly improved tumor targeting when placed in a diabody format. However, additional improvements are required for effective RIT. The applicant will create an antibody pre-targeted radioimmunotherapy strategy employing scFv fusion proteins targeting HER2/neu and haptens to localize radiometals to tumor sites. This will be accomplished by panning a human scFv phage library to isolate scFv reactive with the chelate, CHX-A." These scFv will be affinity matured and then fused to the C6.5 diabody to create a bispecific fusion protein that can capture systemically administered 9OYttrium chelated to CHX-A" and concentrate the radionuclide at tumor sites. Tumor targeting and preclinical radioimmunotherapy studies will be conducted. The advantages of this pretargeting strategy will be sought by also conducting preclinical radioimmunotherapy studies employing the directly conjugated fusion protein or C6.5 IgG1. These studies will identify candidate molecules for clinical development, and will provide a foundation for the clinically effective radioimmunotherapy of solid tumors.